Beruflich Dokumente
Kultur Dokumente
by
Piet SERCU
Raman UPPAL
D/1993/2376/11
Negotiation, Valuation, and Tax Planning
an Integrated Approach
By
and
an Integrated Approach
The standard ("domestic") investment analysis procedure is to evaluate the project in two steps,
first focusing on the operational aspects, and only afterwards dealing with the financing side
(see e.g. Brealey and Myers (1991). International problems give rise to many additional issues,
including interactions with the company's other businesses, political risks, the effect of
exchange risk and capital market segmentation on the required rate of return, and international
taxation aspects of the remittance policy. This last aspect is caused by the fact that most foreign
ventures are carried out via a separate subsidiary, whereas 'domestic' projects are typically
assumed to be in-house so that the 'parent' is immediately and automatically the full owner of
the project's cashflows. Refining a suggestion by Shapiro (1985), international capital
budgeting can, therefore, be implemented in three steps: (1) the "bundled" (branch) valuation,
where the focus is on the economics (operations, exchange and political risks, etc.), and where
the project is assumed to be carried out within a foreign branch; (2) the "unbundling" stage,
where the tax implications and political risk aspects of various intra-company financial
arrangements are considered; and finally, (3) the adjustments for the effects of external
financing.
Another complication arises from the fact that many foreign projects take the form of a joint
venture (N). For a JV, we can essentially follow the three-step approach outlined for a WOS.
That is, we can first do the NPV exercise assuming the foreign business is conducted as a joint
branch, i.e. an unincorporated where all profits and losses are shared pro rata by the partners.
Also the financing of working capital, and the funds resulting from depreciation tax shields
As suggested in the last point, investment analysis for a joint venture is inherently more
complicated than an evaluation of a WOS project. For one thing, NPV-analysis is mixed with
the issue of profit-sharing and is, therefore, hard to separate from the contract negotiation
itself. To further complicate matters, the valuation of a given cashflow by one partner may
differ from the valuation by the other partner(s) because of, for instance, different tax rules, or
because of heterogenous required returns reflecting capital market segmentation. In this article,
we will discuss a framework to deal simultaneoulsy with these issues. In Section 1 we review
the basic bargaining theory. An application for the simplest possible pure-equity joint branch is
provided in Section 2. In Section 3 we explain how issues like differential tax treatment, or
heterogenous required returns can be built in. Sections 4 and 5 deal with two possible
departures from the pure-equity framework: royalties, and capitalisation of the value of the
know-how (brought in as 'equity in kind'). Section 6 concludes the paper.
1. In practice, the joint branch's working capital needs can be financed through bank loans. But as each parent
implicitly guarantees the joint branch's loans, this is just one form of outside borrowing by the parents. If there
were no costs or tax issues, such borrowing would not affect the NPV; and if there are borrowing costs and tax
benefits, these should be considered in Step 3. Likewise, the joint branch can invest free cash flows in the
money market. But in the absence of tax effects or transaction costs this does not affect the NPV; and
transaction costs or tax effects should be part of Step 3.
2. The same advantage is, however, offered by a pure-equity JV.
The first question is what the total NPV is if the bargaining leads to a JV. Denote this as
NPVJv, and assume initially that this is independent of how the JV splits the cashflows from
the operations. (That is, for the time being we assume there are no differential tax effects, nor
heterogeneous discount rates; we shall consider all this later on). The next question is what the
parties get if the negotiations break down. For instance, each company may then pursue the
investment on its own, or one or even both of them may abandon the idea altogether. Denote
the values of these alternatives as NPV A and NPVB, respectively. Obviously, no party will
accept a JV contract that gives it less than the NPV it can realise on its own. One immediate
corollary is that JV negotiations will work only if there are synergy gains, i.e. if NPV JV >
NPV A+ NPVB. Suppm;e there is a positive synergy, NPVJv- (NPVA + NPVB) > 0. Since
each party wants at least its 'own' NPV, the discussion is about the division of these synergy
gains. The usual rule of thumb in practice (which can be justified theoretically; see Rubinstein
(1982), Sutton (1986)) is to split the gains equally) Thus
The simplest way to implement this is a pure-equity contract, under which A and B agree to
share both the costs and benefits on a the same prorata basis. The third and last ingredient of a
bargaining game is provided by outside options. One partner, or both of them, may get offers
from third parties (i.e. companies not explicitly involved in the negotiations). Obviously,
bargaining must also lead to an outcome that is at least as good as the best outside option
available to each player. We now apply this framework to some generic cases with varying
degrees of complexity. The focus is on the implementation of the split-the-difference rule; that
is, we will not repeat the point about outside options.2
1. If the bargaining strengths are nor equal, the synergy gain will be split differently. In what follows, we
assume equal bargaining strengths; it is straightforward to adjust the procedure for a different sharing rule.
2. In fact, the outside options can be huilt into the NPV A and .:-.fPVB figures, as suggested by Sutton (1988).
For instance, suppose NPV A= 100, NPVB=lOO, NPVJv=250. Without outside options, A and B should get
Clearly, A will never accept a deal where its share in the JV's NPV would be below NPV A=
152, because otherwise it would prefer the wholly-owned subsidiary (WOS) alternative. The
synergy gain equals an impressive 493 - (152 + 0) = 341. So A and B agree to split the
synergy gain: B gets a NPV of 341/2 = 170.5 (or 35% of the total NPV), and A obtains 152 +
341.5/2 = 322.5 (or 65% of the total NPV). Under the pure-equity solution, they each shoulder
the corresponding fraction of the investment, and likewise share all cashflows on a 35/65 basis.
3. Case II: A Joint Branch where the JV's value Depends on the
Profit-Sharing Agreement.
The analysis becomes somewhat more involved if the value of the JV depends on the contract
itself. For instance, A may be taxed more heavily than B, because of a higher domestic tax rate
and a credit tax system. In such a case the value of the JV is higher the higher B's stake,
because this reduces the overall tax burden. A similar problem arises if, in segmented capital
markets, the two prospective parents use different discount rates. For example, if the host
country has a closed capital market, an outsider like A is likely to require lower returns than B's
shareholders, because many risks that cannot be diversified away by A's shareholders are
undiversifiable from the point of view of B's shareholders. In this case, the value of the JV is
higher the higher the share that accrues to A.
125 each. If company A has an outside option worth 130, it should get at least 130. Sutton compute B's share
as 130 + (250- (130+100))/2 = 140.
1. A fuller version of this paper, available on request, has detailed cash-flow examples and details about
working-capital adjustments. But all this detracts from the main analysis.
Recall that we are considering a joint branch where all partners fully share pre-set fractions of
all cashflows. Denoting A's share in the investment and cashflows by <j>, the NPV realized by A
equals its share in the before-tax operating cashflows, minus the taxes paid by A, and minus
A's share in the initial investment:
This is just <1> times the NPV JV computed as if there were a 40% tax rate. Let us denote this
figure by NPVJV,40% By a similar argument, B gets (1-<j>) times NPVJV,35%, where in
NPVJV,35% all taxes are computed using a 35% tax rate. The aggregate NPV value then is
This yields
<j> X
465 = 135 + (493- 2~ <j>]- 135
or
<1> (465 + 228 ) = 135 + 493 2135 314
=314 =><1>=465+14=65.5%
1. In integrated capital markets, the present value of a cashflow in A's currency (using currency-A expected
cashflows and a currency-A discount rate) must be the same as the present value of the cashflow expressed in
currency B and discounted at a currency-B discount rate, multiplied by the current spot rate. So the valuation
currency is not an issue.
A similar approach can be adopted if the partners use different required returns. Suppose that
partner A requires a 20% return, and partner B a 25% rate. Denote by NPV JV,.20 the NPV of
the entire cashlow discounted at 20%, and by NPV JV,.25 the result if discounting is at 25%. If
company A gets a fraction <P of the cashflows in return for a fraction <P of the investment, it can
compute the NPV of its share in the benefits and costs as <P x NPVJV,.2o; likewise, B's share in
the benefits and costs is (1-<j>), so that the NPV of its part equals (1-$) x NPVJv,.25 The total
value again depends on (j), in a linear way:
We just bring in this formula into the sharing rule. Company A should get
NPV
tt-.
'!'X JV,.20 = NPV A+ NPVJv- (NPV A+ NPVB)
2
For instance, if NPVJV,.25 = 360, NPVJV,.20 = 493, NPVA = 152, and NPVB = 0, we find
(j} X
493 = 152 + (360 + 1~3 (j}] - 152
or <P = 4932~~ 6 . 5 = 60%. Thus, B's share goes up from 35% (base case) to 40%. This is
the result of its compensation for its lack of diversification opportunities. If NPVB had not been
equal to zero, B's starting position would simultaneously have been weakened by its higher
discount rate, which would have attenuated the compensation effect.
* * *
Thus far, we only considered pure-equity joint-branches, where profits and other cashflows are
shared pro rata of the fractions in the investments. In Step 2 of the NPV-procedure, the joint
operation becomes an incorporated business, and non-proportional sharing arrangements
become possible. I The JV's NPV can now be shared in various ways: for instance, there may
1. In addition, the JV can now borrow from outside sources. But this is part of Step 3. In Step 2, we just
consider arrangements betwen the parents and the JV.
A similar approach can be adopted if the partners use different required returns. Suppose that
partner A requires a 20% return, and partner B a 25% rate. Denote by NPV JV,.20 the NPV of
the entire cashlow discounted at 20%, and by NPVJV,.25 the result if discounting is at 25%. If
company A gets a fraction <1> of the cashflows in return for a fraction <1> of the investment, it can
compute the NPV of its share in the benefits and costs as <1> x NPVJV,.2o; likewise, B's share in
the benefits and costs is (1-<!>), so that the NPV of its part equals (1-<1>) x NPVJv,.25 The total
value again depends on <j>, in a linear way:
We just bring in this formula into the sharing rule. Company A should get
NPV
Jt.
'i' X JV ,.20 = NPV A + NPVJv - (NPV A+ NPVB)
2
For instance, if NPVJV,.25 = 360, NPVJV,.20 = 493, NPVA = 152, and NPVB = 0, we find
2
or <I>= 493 ~~ 6 . 5 = 60%. Thus, B's share goes up from 35% (base case) to 40%. This is
the result of its compensation for its lack of diversification opportunities. If NPVB had not been
equal to zero, B's starting position would simultaneously have been weakened by its higher
discount rate, which would have attenuated the compensation effect
* * *
Thus far, we only considered pure-equity joint-branches, where profits and other cashflows are
shared pro rata of the fractions in the investments. In Step 2 of the NPV -procedure, the joint
operation becomes an incorporated business, and non-proportional sharing arrangements
become possible. I The N's NPV can now be shared in various ways: for instance, there may
1. In addition, the JV can now borrow from outside sources. But this is part of Step 3. In Step 2, we just
consider arrangements betwen the parents and the JV.
Our approach is to chose p such that, faced with the exogenously fixed <!>=.49, company A still
gets half of the synergy gains. What A gets is the after-tax royalties, i.e. p x salest x (1-.35),
plus 49% of whatever is left after paying out the royalty; while B gets 51% of the JV's
cashflows after royalties. If, as we are assuming, taxes are fully neutral, the total Gross PV
(GPV) is unaffected:
We now set p such that each partner gets the same NPV as under the joint-branch solution;
synergy
company A should get 322.5 (= NPV A+ = 152 + 341/2):
2
In the above example, we fixed <1> at .49, and did not use any upfront payment for the
knowhow (i.e. we set K equal to 0); so the free parameter was the royalty percentage, p. In the
next example, we shall set <1> and p on the basis of exogenous considerations, and solve for the
upfront compensation for know-how K that achieves the desired division of the synergy gain.
Such an upfront compensation can come in at least two guises. First, it could be stuctured as
part of a licensing contract: the amount K is paid out to one of the partners as an upfront
licensing fee. This structure means that K is surely taxable income to the recipient (company
A). Alternatively, A can bring in its know-how as equity 'in kind'; that is, both partners agree
that the knowhow is worth K, and list it as a non-tangible, depreciable asset on the balance
sheet. On the liability side of the balance sheet, equity consists of the part put up in cash,
denoted as CashA + Cashs, and the part brought in in kind, K. Partner A, who brings in the
know-how, would then get a part of the profits equal to <1> = (K + CashA)I(K + CashA +
Cashs). The JV can depreciate the intangible asset, and get the corresponding tax shield. But
company A may be able to book K as an unrealised capital gain, in which case it would escape
taxation for the time being. Thus, there is a tax advantage relative to the ftrst variant. And even
if there is no corporate tax advantage for company A, there may be an advantage in terms of
withholding taxes because, formally, no income is paid out.
To illustrate aN contract of this type, let us ftx p at 5%, for instance because any amount
higher than that would be deemed "not at arms' length" by the host country taxman. This is
lower than the 8.25% that achieves the desired division of the JV's NPV, so that A wants
another form of compensation. Under the proposed scheme, A brings in know-how valued at
K, plus possibly some cash, and B brings in some cash. Total cash injections (I) are the same
The amount K is capitalized, and depreciated linearly over five years by the JV, but by
assumption A is not taxed on K as there is no realised capital gain. This is obviously interesting
from a tax point of view: there is a tax shield, discounted at the JV's borrowing rate (say 17%)
and worth Lt=1,5 (.K/5 x .35)/1.17t+1 = K x .191.1 On the other hand, the amount K is not paid
up in cash, so at time 7 (when the JV is liquidated) there will be a tax when K is paid out as part
of the liquidation dividend. Thus, taxes are postponed rather than avoided, and the net saving is
reduced tp K x (.191- .35/(1.17)7 = K x .075.
Our problem now is to find a fair value for K. The basic principle is the same as before: B and
A still split the synergy gains, with
From A's point of view, we have to identify CashA (and, by implication, K), such that its share
in the NPV satisfies the split-the-difference rule:
1
.49 X 493 +.51 .05 (1-.35) 2962 +.51 X K = 152 + 2 [493 + K X .075- 152]
which yields
1. Taxes are assumed to be paid in the middle of the year following the reporting . .-ar, hence the "t+ 1" exponent
in the discounting.
The priciples used in these simple examples allow one to tackle problems where many
complications arise simultaneously. Always start from the split-the-difference rule, write the
N's NPV and the partners' shares as a function of the predetermined parameters and the free
parameter to be identified. Some iterations may be needed if the pre-set parameters imply
unexpected solutions for the residual parameter-like a value for <1> that is not between 0 and 1.
Adler, Michael and Bernard Dumas. "International Portfolio Choice and Corporation Finance:
A Synthesis," Journal of Finance, June 1983, 925-984.
Lessard, D., "Evaluating Foreign Projects: An Adjusted Present Value Approach," in Donald
R. Lessard (ed), International Financial Management, Theory and Application, Wiley,
1985.
----------- and Alan Shapiro. " A Framework for Global Financing Choices," working paper,
Sloan School of Management, MIT, October 1982, p. 17.
Miller, Merton H. "Debt and Taxes," Journal of Finance, 1977, v32(2), 261-275.
-----------. Multinational Financial Management, Allyn and Bacon, Needham Heights, Mass.,
1991.
Rubinstein, A., Perfect Equilibrium in a Bargaining Model, Econometrica 50(1), 1982, 97-109